Only if you use flawed metrics. If you use the same metrics for health and the same methodology of a reasonable scientific paper then you get a purely better result. the gotcha is that it is useless for everyone but you.
A problem would be if you do too many experiments too often, but this you also need to control in n=50.
There is a huge difference between subjectivity and n=1
I don't understand how this is feasible. As a young healthy male I never had my full blood count have values anything other than within reference range. Even when I ate noodles everyday in college because I was broke, or now I eat balanced and healthy unprocessed food, veggies, lean meat etc I have similar blood counts. How do you measure a diet's effect within your lifetime? It seems like even if you look healthy by all measures a normal person is capable of measuring right now, but still be eating unhealthy.
> but still be eating unhealthy.
but maybe you are at a good health level while you only eat junk food. Like people who smoke all of their life and have no health issue.
Diet researches do not measure lifespans of participants.
Those who do are rare, as they would need to last at least a few decades. IIRC to publish actual research on diet effect on lifespan you need 80 years.
What you do is develop metrics useful to predict lifespan and quality of life and then see the impact of a diet of those metrics. There is nothing you cannot do privately here (using some laboratory for analysis obviously)
I am not saying it is easy and for sure it is error prone, but so is science itself.
A problem would be if you do too many experiments too often, but this you also need to control in n=50.
There is a huge difference between subjectivity and n=1